More waves in longitudinal studies do not help against initial sampling error
Suppose you ran 20 studies. By coincidence, one of them attains p < .05 and, by good fortune, it's in the right direction too. The other studies get thoroughly examined…
James Heckman